localized within the mesangium of both groups (Figs. 3a and c). However, localization of IgG was much more diffuse in Group I rats and also involved.
ENHANCEMENT DEPOSITION
OF
GLOMERULAR
BY A CIRCULATING
IMMUNE
COMPLEX
POLYCATION
BY JEFFREY L. BARNES AND MANJERI A. VENKATACHALAM From the Departments of Pathology and Medicine, The University of Texas Health Science Center, at San Antonio, San Antonio, Texas 78284
T h e net charge o f antigens, antibodies, or o f i m m u n e complexes may affect their localization in glomeruli (1-3). A positive net charge favors increased localization in the g l o m e r u l a r basement m e m b r a n e (GBM), 1 whereas a negative net charge favors mesangial deposits (1-3). T h e s e studies e x a m i n e d the interaction o f differently charged i m m u n e reactants with polyanions ~ in normal giomeruli. Conversely, perturbations o f glomerular polyanions may alter glomerular interactions with i m m u n e reactants and affect their localization. Hypothetically, secretory cationic proteins from activated ieukocytes and platelets may bind to glomeruli d u r i n g the earliest stages o f i m m u n e injury and enhance glomerular permeability, favoring i m m u n e complex deposition. Circulating exogenously derived polycations bind to polyanions in the GBM and increase permeability to proteins (6-9). H e r e , we show that administration o f the polycation polyethyleneimine (PEI), which is known to bind to the GBM in vivo (10) markedly enhances the deposition o f p r e f o r m e d i m m u n e complexes in glomeruli, particularly in the GBM. Methods Immunizations. New Zealand White rabbits were immunized against bovine serum albumin (BSA) Fraction V (11). Sera were assayed by immunodiffusion, and the strongest antisera were pooled. Pooled serum was fractionated by precipitation with 18% saturated sodium sulfate, and dialyzed against 0.02 M phosphate-buffered saline, pH 7.35 (PBS). Antibody was assayed by quantitative immunoprecipitin analysis (12). Isoelectric focusing of the immunoglobulin fractions on slab gels of 5.0% acrylamide and 1.0% bis-acrylamide containing 2.5% ampholine (pH range 3.5-10.0) revealed multiple bands ranging in isoelectric points (pI) between 5.3 to 8.0. The pI of BSA was determined to be 4.9-5.1. Preparation of Immune Complexes. Soluble immune complexes were prepared as described by Germuth et al. (13). The equivalence amount of BSA was added to pooled This investigation was supported by research grants AM 17387 and AM 30393 from the National Institutes of Health. Send reprint requests to: Jeffrey L. Barnes, Dept. of Pathology, The University of Texas, Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78284. ' Abbreviations used in the paper: BSA, bovine serum albumin; FITC, fluorescein isothiocyanate; GBM, glomerular basement membrane; PBS, phosphate-buffered saline; PEI, polyethyleneimine. 2 Connective tissue elements with a net negative charge are known to be present in the cell surface coats of glomerular capillary endothelial cells and epithelial cells, and in the GBM. Such polyanions, which may be sialoproteins, protein carboxyls, and sulfated proteoglycans, are thought to play major roles in many aspects of glomerular structure and function. Polyanions in the GBM, in particular, appear to determine charge selective interactions of the glomerular capillary wall with macromolecules (4, 5). 286
J. ExP. MED.© The Rockefeller University Press • 0022-1007/84/07/0286/08 $1.00 Volume 160 July 1984 286-293
BARNES AND VENKATACHALAM
287
fractions, mixed, and incubated at 37°C for I h and 4°C overnight. The precipitate was washed three times with PBS, then suspended in a BSA solution to make soluble immune complexes at 40 times antigen excess and incubated at 37°C for 1 h and 4°C overnight. Before injection into rats, the immune complex solution was centrifuged at 48,000 g for 60 min and sterile filtered. In additional studies, affinity-purified anti-BSA was labeled with ~25I (Iodine) and then added to the anti-BSA immunoglobulin fraction before preparation of immune complexes as outlined above. Polycation. PEI, 18 (mol wt 1,800), obtained from Arsynco, Inc. (Carlstadt, NJ), was diluted in 0.05 M PBS, pH 7.2, to make a final solution containing 0.5% PEI with a pH of 7.4. In control experiments, 0.05 M PBS, pH 7.4 without PEI was used. Protocol. 12 male Sprague-Dawley rats (Harlan Sprague-Dawley, Madison, WI), weighing 200-290 g, were paired according to age and size, then divided into two groups. Rats from Group 1 were briefly anesthetized with ether and then injected with PEI at a dose of 10 ug/g body weight via a tail vein. PEI administered in this manner binds to polyanions present in the GBM and enhances glomerular permeability to the tracer macromolecule ferritin (molecular radius -- 61A (8). 5 rain later, ~ 1 ml of preformed immune complexes containing 30 mg of anti-BSA antibody in 40 times antigen excess was injected intravenously. Controls (Group II) received PBS without PEI followed by the same dose of immune complexes as described above. To assess any toxic effect of PEI, an additional four rats (Group III) were given PEI, as outlined above, but without subsequent injection of immune complexes. In each group, all injections were repeated two additional times 4 h apart. After each series of injections, the rats were placed in metabolic cages and urine collected for subsequent protein analysis. 1 h after the last injection, the rats were anesthetized, the left renal pedicle ligated, and the left kidney excised, sliced, and frozen in chilled isopentane. Immediately following removal of the left kidney, the right kidney was fixed by perfusion with 1.25% glutaraldehyde in cacodylate buffer, pH 7.4, and subsequently processed for light and electron microscopy using routine methods. In an additional study, the ultrastructural localization of immune deposits was assessed by indirect immunoperoxidase procedures to reveal rabbit IgG. Immune reactions were performed on 60-#m tissue sections and exposure time to 3,3'-diaminobenzidine was kept to a minimum (1 rain) as recommended by Courtoy et al. (14). Glomerular localization of BSA and rabbit IgG was assessed in sections of renal cortex by direct immunofluorescence, using fluorescein isothiocyanate (FITC)-conjugated IgG fractions of appropriate antiserum. Frozen sections (8-#m thick) were viewed and photographed using a Zeiss Universal II Research microscope equipped with epifluorescence and using a standard FITC filter set supplemented with a KP 560 shortwave pass filter. The intensity of glomerular immunofluorescence of localized BSA and rabbit IgG was quantitated by photometric analysis, using a Zeiss photometer attachment SF connected to a PMI-2 indicator. Intensity of fluorescence in a 125-gm field (approximate size of glomeruli) was obtained by using a 2.5-mm aperture in the light path between the specimen and the photometer. Fluorescence intensity of glomeruli was determined as the relative percent transmission compared to a fluorescein filter standard. All comparisons were expressed as the ratio of fluorescence intensity between PEI (Group I) and control (Group II) in each paired experiment. At least 25 glomeruli were examined in each tissue section. In experiments in which radiolabeled immune complexes were used, blood samples were collected 5 min before the third injection of PEI or PBS and at termination of the experiment, to determine the serum concentrations of ~2~I-immune complexes. The size distribution of immune complexes in serum samples was determined by sucrose density ultracentrifugation. 200 #! of serum was applied to sucrose gradients (10-40%), followed by centrifugation at 100,000 g for 16 h. Size distribution of immune complexes was determined by continuous spectrophotometric measurements at 280 nm and by isotopic analysis of 15-drop fractions. Results Administration o f PEI e n h a n c e d the deposition a n d altered the distribution o f g l o m e r u l a r deposits as d e t e r m i n e d by electron microscopy. All capillary walls
288 ENHANCEMENT OF IMMUNE COMPLEX DEPOSITION BY A POLYCATION from all glomeruli examined of Group I, showed numerous small electron-dense deposits in the subendothelium and to a lesser extent in the subepithelium (Fig. 1 a). Peripheral capillary wall deposits were absent in glomeruli of Group II (controls) (Fig 1 b). T h e mesangial matrix of both groups showed infrequent dense deposits by electron microscopy. Deposits were more numerous in Group I rats than in controls (Group II). Immunoperoxidase methodology revealed the localization of rabbit IgG identical in distribution to that of the dense deposits observed by routine staining techniques, substantiating their immune nature (Figs. 2 a and b). Hypercellularity or inflammatory cell infiltrates were not observed in any of the three groups. Some giomeruli from Group I showed luminal strands of fibrin and platelets; however, this observation was focal and most capillary loops were patent and free of luminal material. Glomeruli of controls (Group II) did not show fibrin or platelets in capillary lumina. Similarly, administration of PEI alone without subsequent injection of immune complexes (Group III) did not result in glomerular alterations. Enhancement of glomerular localization of immune complexes by PEI was verified by quantitative immunofluorescence. Glomeruli of Group I rats showed substantially greater intensity of fluorescence of localized rabbit IgG and BSA compared with paired controls (Group II) (Table I). Rabbit IgG and BSA localized within the mesangium of both groups (Figs. 3a and c). However, localization of IgG was much more diffuse in Group I rats and also involved some peripheral capillary loops (Fig. 3 a). Despite localization of immune deposits
FIGURE 1. Electron micrographs of glomerular capillary walls of (a) Group I (PEI plus immune complexes) and (b) Group II (control, diluent plus immune complexes). Electrondense deposits (arrows) are localized in the subendothelial and subepithelial aspects of the GBM of Group I (a), but absent in the GBM of Group II (b). × 37,000. FICURE 2. Glomerular localization of rabbit IgG by immunoperoxidase electron microscopy. In glomeruli of Group I (a), reaction product (arrows) is localized in an identical distribution as the dense deposits observed by routine staining methods (see Fig. 1 a). Reaction product (rabbit IgG) is absent in the GBM of Group II controls (b). × 49,000.
BARNES AND VENKATACHALAM
289
FIGURE 3. Glomerular immunofluorescencelocalizationof rabbit IgG of Group ! (a and b) and Group II (c) rats. In Group I, staining is predominantly mesangial but some peripheral capillary loops (arrows) also show IgG localization(a). In some rats localizationof rabbit IgG was observed in peripheral capillary loops (b). Group II controls showed only mesangial localizationof rabbit IgG in a segmental distribution (c). x 340. in the GBM of all peripheral capillary loops by electron microscopy (Figs. 1 a and 2 a), IgG and BSA were not always detected by immunofluorescence in this location, most likely reflecting the lower sensitivity of light microscopic techniques. However, in two Group I rats, rabbit IgG and BSA were also observed in a linear pattern in glomerular capillary loops (Fig. 3 b). In all cases, localization of rabbit IgG and BSA was absent in peripheral capillary loops of glomeruli from Group II rats, corresponding to the electron microscopic observations. Serum concentration of l~5I-anti BSA-BSA i m m u n e complexes at the first collection period was not significantly different between Groups I and II (Group I = 2.59 + 0.25 (SEM) m g / m l vs. G r o u p II = 2.12 + 0.25 m g / m l (P < 0.20). However, serum concentration o f immune complexes in blood samples taken at the termination of the experiment was significantly greater in Group I (4.50 + 0.27 mg/ml) when compared with Group II controls (3.66 _+ 0.17 mg/mi, P
